# voltage regulator question

Discussion in 'Electronic Basics' started by tempus fugit, Dec 1, 2003.

1. ### tempus fugitGuest

Hey all;

My understanding of a voltage regulator is that it is supposed to deliver
(in this case) 15v regardless of the current draw (within spec of course).
However, if I drop a 2K resistor to ground on my homebrew PS, I notice that
the voltage drops about 10-15 mV. Or am I off base - is it only supposed to
deliver 15v regardless of line voltage (assuming the line voltage is above
the dropout voltage of the regulator)?

Put another way, if I'm looking at the voltages on a piece of gear (could be
anything) that's using say, a regulated +15v and -15v, will the voltages at
the regulator read +15 or -15v with the unit on? I.e., if the unit is
drawing some current, but not beyond what the regulator is spec'd for, will
there be a corresponding voltage drop, or is the regulator supposed to keep
the voltage at +15v irregardless?

Thanks

2. ### CFoley1064Guest

Hey all;
Did you use a 7815, an LM723, or something else? Linear or switching

The term you're looking for is "load regulation", and is usually defined in
terms of percent (of nominal voltage) from no load to rated load, or from
minimum to maximum load. It states that the regulated voltage will not vary

For a small switching power supply, 15 mv drop (0.1%) from no load to 7 mA
wouldn't be unusual at all -- actually, I'd expect at least that much. For a
linear power supply, it might mean something's wrong, but not necessarily. You
might want to try different loads on the power supply, and seeing if it's
relatively stable over a wider load range. If the load regulation stabilizes
and remains less than rated spec, you're probably OK. If not, check your
wiring and layout first.

Good luck.
Chris

3. ### John LarkinGuest

I've seen (and built) several switchers that had negative regulation:
the higher the load current, the higher the output voltage. I've never
really analyzed why.

John

4. ### tempus fugitGuest

Sorry Chris - it is a variable dual polarity supply using a 317 and a 337

5. ### cpemmaGuest

If you've used high-value resistors in the R1/R2 control chain you'll see
that drop, check the datasheet for minimum load current.

6. ### defaultGuest

As someone has already said, you don't mention the specifications for

One other thing is where you measure the voltage drop - at the point
of regulation, or at the end of a few feet of wire along with the load
resistor.

If the regulator is "remote sensing" it would boost/cut the output
voltage from the regulator so it would remain within the load
regulation spec at the point of use. (compensates for drop in the
wiring).

The word you want to use is "regardless."

7. ### tempus fugitGuest

OK I tried dropping a 100 ohm resistor to ground and got a voltage drop of
about 2.5v this time (i.e., the voltage went from 15v to 12.5v). The
regulator in question is a 317 (I get the same results with a 337). Both of
these specify a 0.1% load regulation. Obviously, the 2.5v drop is a tad
higher than 0.1%.
Am I testing this properly, or is this what I should be seeing?

Thanks

8. ### defaultGuest

Right on! Well outside of specification for the part (assuming the
wire resistance is low and you measured at the regulator).

Given that the wire length and connections are good:

The logical explanations are bad part(s) or too high ripple (not
enough filtering) or poor transformer regulation.

What is the input to the regulator under load? It has to be 4 volts
higher than the output voltage. That is four volts to the "valley" of
the ripple. (4 volts is typical for a LM317)

Some generalizations: Transformer regulation is usually (but not
always) better than 10% no load to full load. Big heavy transformers
usually have better regulation (at low currents especially).

A good rule of thumb for filter caps is 1,000 micro farads per 1/2 amp
of current, for about 10% ripple. (electrolytic caps can be specified
at -10% to +125% so you might check that spec as well)

You must be overlooking something.

10. ### tempus fugitGuest

Thanks Default. I had a few minutes to check some things out tonite. The
input to the regulator is about 18v, which is cutting it a little close for
15v, but I get similar results when it is set to 5v. I scoped the input to
the regulator; there is no ripple here (I'm using 4700uF caps).

I'll double check the wiring and make sure everything's OK there too.

Thanks again

11. ### tempus fugitGuest

I found the problem. I had put 2 8 ohm resistors in series with the output
to protect from accidental shorts. This would keep the maximum current draw
at a little less then an amp (cutting it close I know). When I tested it
without the resistors, it gave me a consistent output voltage.

Of course, I now have no short circuit protection. Any ideas for something
simple to provide that? Also, why did the 16 ohms in series with the output
stop the regulator from regulating?
Thanks

12. ### defaultGuest

Well that would certainly do it. Putting 16 ohms in series with the
output drops voltage.

The water pipe analogy: pipe (resistor) is too small for the flow
(current) and pressure builds up on the inlet side and drops on the
outlet side (voltage drop across the resistor).

If I'm not mistaken, the LM317 adjustable three terminal regulator has
built in over current protection at something like 1.5 amps (will
protect itself and the transformer if the transformer can output 1.5
amps safely. Voltage drops as current tries to increase beyond its

The LM317 will get warm or hot while limiting and it also includes
thermal protection, so the output current maximum will come down as
the part gets hot.

If you want some other current limit (less than the 1.5 amps the part
is good for) you'd have to add an external resistor to sense the
current and shut down the regulator - but then the design gets tricky
and you loose the ease of application of the 3 terminal regulator.

I generally will use a TO92 size regulator if I want to keep the
current low; they limit at 100 milliamps. If I need something else I
go for the LM723/UA723 (14 pin dip package) and external pass
transistors - that allows a current limit to be set with an external
resistor while keeping the resistor inside the control loop (doesn't
change the regulation - the resistor only need drop .6 volts to limit
the current)

The databook (National Semiconductor Voltage Regulator Handbook) shows
a current limited battery charger. That takes an additional
transistor and three resistors and one resistor is in the ground leg
in series with the load (battery being charged). They show an
adjustable current regulator using a 3 terminal part (LM338 - 5 amp
regulator so the LM317 would also work) to achieve adjustment they add
another LM117 to handle the current regulation function.

You don't say that you need the current to be adjustable just want
short protection - the LM317 already does that.

The National Semiconductor web site has the application notes on it
and they will ship you a CD of their web site with all the parts they
make for a nominal shipping cost or free. The CD is searchable just
like the web site if you use a browser to open it. The regulator
handbook was also free, but published in 1982 and is probably out of
print now.

13. ### tempus fugitGuest

"> Well that would certainly do it. Putting 16 ohms in series with the
OK, but I figured that there would already be a certain resistance between
the output and ground (say, 100 ohms) and adding a little more wouldn't
change things much, particularly when it was much smaller than the circuit's
resistance (i.e., If it was 100 ohms at 10v, the current draw would be 100
mA, so it would be a little less at 116 ohms). Also, the voltage drop varied
with the current draw - it was pretty major when drawing 100mA, but only a
few 100 mv whn drawing 30 or 40 mA. Why would that be?

That's good news. I'll check the datasheet and see if it does.
No, I don't really need adjustable current, just wanted something to prevent
me from having to replace the regulator IC every time I accidentally touched
the hot out to ground.
Thanks again

14. ### defaultGuest

I don't know why that would be, but the first thing I'd look at would
be the input voltage with a meter or better yet a scope to see what it
looked like. You have to maintain the 4 volt differential between in
and out or the part drops out of regulation. The meter averages the
voltage so it won't tell you if there is too much ripple on the input.

"Low Drop Out" regulators will run with ~.6 to 1.0 volt differential.
(makes them more efficient and less heat to dissipate - if the input
voltage is lower with respect to the output)

You may have a problem with too much ripple, bad connections, or just
a ground loop causing an oscillation. The three terminal regs are
easy to apply, but they can be made to oscillate if there's a ground
loop or too much inductance on the output.

Without a scope you won't see some of the things that can go wrong.

Oscillations are not common, at least not with a good layout . . .
Most designers probably never encounter an oscillation in a 317. A
small output cap can do wonders if the load is inductive and causing
an oscillation. For a layout problem, the layout has to be corrected.
Three terminal regs are as close to bullet proof as any part is. You
can destroy them by putting a reverse current through them by
connecting the input backwards, or having a large capacitor on the
output and dropping the input voltage suddenly (like discharging the
filter cap with a screw driver) Once they are wired correctly and in
use they last a long time.

15. ### tempus fugitGuest

Thanks again Default.

I checked the data sheet, and it says that the typical current limit on the
TO220 (which is what I'm using) is 2.2A. Wouldn't that destroy the regulator
(which is spec'd for 1.5A) in a short circuit situation? Or am I looking at
the wrong spec (or interpreting it wrong)?

Thanks again.

16. ### defaultGuest

The part is spec'd with a minimum of 1.5, typical of 2.2, and maximum
of 3.7.

The one you get will fall in that range. If you are a commercial
designer you use the minimum current, because they guarantee the part
to work at 1.5, the other currents are possible and most of the parts
they turn out are 2.2.

If you are in a position where you just bought 100,000 three terminal
regulators for a product - they don't want you sending them back if
they only output 1.5 - and they don't want to select parts for a
particular user . . .

That doesn't mean the regulator will be destroyed at 2.2 amps out.
With a good heat sink it should put out 2.2 amps or even 3.7 without
complaining. But if it only puts out 1.5 before it current limits -
it is still a good part.

The parts tend to get better with manufacturing experience, so it may
be that very few will output only 1.5 these days - but if you are
committing large amounts of money to a product, you would be wise to
observe the lower limit or call National and see what they are willing
to do for you.

From your hobiest/experimentor perspective, you don't have to worry
about the part, it is protected from over current and shorts.

You still need to worry about the power source. If you have a
transformer supplying power for the device it must be able to source
3.7 amps to the regulator without burning up. (that could actually be
higher - like 5 amps - depending on how the transformer is specified)

For arguments sake, you have a transformer that will put out 12 volts,
at 3 amps. That is 12 volts RMS AC voltage at 5 amperes. Use a full
wave bridge rectifier to filter it and you might have 17 volts of DC
(1.4142 * the RMS voltage, give or take a little for transformer
regulation and rectifier losses)

That doesn't mean you get to suck 17 volts at 3 amps from the filter
cap side of things. 3 amps/17 volts at the filter is close to 4.25
/12 at the transformer.

If your only concern is that the regulator and transformer survive a
temporary short, don't sweat it. If you are building something like a
battery charger where the over-current condition can last for hours it
is a different matter.

I learned that lesson a long time ago. Wanted some tunes at work and
built an amplifier, but didn't consider the difference in AC and DC
amps. My power transformer was sealed in a steel case and potted in
tar. I fell asleep one night listening to music in the darkened shop.
The transformer exploded and woke me up.

I also learned something from this discussion. I was looking at class
A audio amps and it occurred to me that one could be made by using a
single output transistor working against a constant current source
(for a push-pull output). That was a year ago - but I didn't try it.
Looking at the old regulator book I saw where they actually do it.
Neat idea.

17. ### defaultGuest

For arguments sake, you have a transformer that will put out 12 volts,
at 3 amps. That is 12 volts RMS AC voltage at 5 amperes. Use a full
wave bridge rectifier to filter it and you might have 17 volts of DC
(1.4142 * the RMS voltage, give or take a little for transformer
regulation and rectifier losses)

That is 12 volts RMS AC voltage at 3 amperes (not 5)

18. ### cpemmaGuest

The L200 adjustable regulator is a 5-pin 2A device, very easy to fit current
limiting (1 extra resistor) if you need to protect the transformer.

On your voltage drop, you produced a simple potential divider - 15v over 16R
+ 200R gives 15*200/216 =13.9 volts over the 200R.

19. ### tempus fugitGuest

I wasn't sure if there was a difference between AC and DC amps. So do you
multiply it by 1.414 just like RMS voltage? (3A*1.414=4.25A)?

That's really my only concern; I use it as a bench supply for prototyping
or repair. I'm sure that I've been careless (stupid?) enough to actually
hook the supply up to a circuit that has a short in it, or (more
embarrassing) plugged the hot and ground into the same connection on a
protoboard. So I guess in this situation the overcurrent situation could
last a minute or so before I caught it. As long as it will withstand
something like this I'm happy.
Cool. I'm glad I'm helping you too.

20. ### defaultGuest

I wasn't sure if there was a difference between AC and DC amps. So do you
Yes, except you have it backwards - as the voltage goes up with
rectification and filtering the current allowed on the output goes
down for the same power level at the transformer.

Multiply by the reciprocal of 1.414 - .707